Network Architecture to Identify Spatial Knowledgefor Epidemic
(An application of datamining)
Abstract:
Recent developments in information technology have enabled collection and processing of vast amounts of personal data, business data and spatial data. It has been widely recognized that spatial data analysis capabilities have not kept up with the need for analyzing the increasingly large volumes of geographic data of various themes that are currently being collected and archived. Our study is carried out on the way to provide the mission-goal strategy (requirements) to predict the disaster. The co-location rules of spatial data mining are proved to be appropriate to design nuggets for disaster identification and the state-of-the-art and emerging scientific applications require fast access of large quantities of data. Here both resources and data are often distributed in a wide area networks with components administrated locally and independently, a framework has been suggested for the above. Our contribution in this paper is to design network architecture for disaster identification.
- INTRODUCTION:
Geography is an integrative discipline and geographic data under analysis often span across multiple domains. The complexity of spatial data and geographic problems, together with intrinsic spatial relationships, constitute an enormous challenge to conventional data mining methods and call for both theoretical research and development of new techniques to assist in deriving information from large and heterogeneous spatial datasets. (Han and Kamber 2001; Miller and Han 2001; Gahegan and Brodaric 2002).. ‘health’ maps have become available as the use of geographical information systems in health related contexts increased.
A formula implemented as Hazard science to Risk Science, towards understanding the hazards and their consequences (risks), following a probabilistic approach using spatial data mining .Due to larger heterogeneity of spatial data, the providers of geographic data specify different models for same spatial objects. Context specific semantics is one of the best approach suggested which deals with provision of feature space derivations. Unknown and unexpected patterns, trends or relationships can hide deep in a huge feature space.
A hypothesis space is formed by all possible configurations of the tools used to detect patterns in a feature space. Characteristically, however, the hypothesis space for a large and high dimensional geographic dataset has an extreme degree of complexity. and make it very hard for analytical methods or visual approaches to find. (Miller and Han 2000).
This is caused by several factors. First, each pattern may involve a different subset of variables from the original data, and the number of such subsets (hereafter subspaces), i.e., possible combinations of attributes, is huge. Second, inside a subspace, potential patterns can be of various forms (e.g., clusters can be various shapes). Third, for a specific pattern form (e.g., cluster of a specific shape), its parameter space is still huge, i.e., there are many ways to configure its parameters. Fourth, patterns can vary over geographic space, i.e., patterns can be different from region to region.
2. APPLYING SPATIAL DATA MINING
Spatial data mining becomes more interesting and important as more spatial data have been accumulated in spatial databases.
2.1. SPATIAL STATISTICS
Using spatial statistics measures, dedicated techniques such as cross k-functions with Monte Carlo simulations, lattice method have been developed to test the collocation of two spatial features. At the outset the studies include, the spatial data mining problem of how to extract a special type of proximity relationship – namely that of distinguishing two clusters of points based on the types of their neighboring features is another study[3]. Classes of features are organized into concept hierarchies [1].A reasonable and rather popular approach to spatial data mining is the use of clustering techniques to analyze the spatial distribution of data. While such techniques are effective and efficient in identifying spatial clusters, they do not support furtheranalysis and discovery of the properties of the clusters.
2.2. MINING COLLOCATION PATTERNS
Mining collocation patterns gives the standard of observing the generic characteristics of a given spatial zone with more relevant boolean features with their s%(support) and c(confidence) [3]. The work of mining Collocation patterns into spatial statistics approaches and combinatorial approaches . The spatial Collocation pattern mining framework presented in the erstwhile works has bias on
popular events. It may miss some highly confident but “infrequent” Collocation rules by using only “support”-based pruning.
In a spatial database S, let F = {f1,…,fk} be a set of boolean spatial features. Let I = {i1,…,in} be a set of n instances in the spatial database S, where each instance is a vector consisting of [instance-id, location, spatial features]. ~ Neighborhood relation R over pair wise locations in S exists ~ is assumed. The object of this collocation rule mining is to find rules in the form of AB, where A and B are subsets of spatial features. A determines the set of spatial features that form the antecedent part of the rule and B defines the action and its consequential parts the support and the confidence. The rule indicates the coincidence of the spatial collocation rule absorbs the action of the rule in the “nearby” regions of the spatial objects that comply with the collocation rule. A collocation patternC is a set of spatial features,i.e., CF. A neighbor-set L is said to be a row instance of collocation pattern C if every feature in C appears in an instance of L, and there exists no proper subset of L does so. We denote all row instances of a collocation pattern C as rowset(C). In other words, rowset(C) is the set of neighbor-sets where spatial features in C collocate. The conditional probability is the probability that a neighbor-set in rowset(A) is a part of a neighbor-set in rowset(B). Intuitively, the conditional probability p indicates that, whenever we observe the occurrences of the spatial features in A, the probability to find the occurrence of B in a nearby region is p
2.3. FINDING/ESTIMATING SYMPTOMS TO BUILD COLLOCATIONS
In the imaginary figure the landscape describes two important spatial marks, sea and lake. The epidemic spread is noted in the Figure-1. The water in the lake is afflicted by the lichens and mosses at the western zone of the lake as most of the water is stagnant and covered by marsh. The people utilizing the water resources at this zone may be affected by so many kinds of fecalcontamination in water and food. The water in the sea is contaminated with the high salts and the crude oil and base products, as the people cannot take the water for the domestic purposes. As it is not easy to detect the spread of virulent features from the spatial data, the features are tested in the affected people who are native of the zone.
Figure-1:Lake affected by lichens & mosses
3. RELATED WORK
A few epidemics that are spread due to common sources like contaminated water and contaminated food are shown below.
Common Source Epidemic DiseasesDisease / Causative Agent / Infection Sources / Reservoirs
Bacillary / Shigella disenteriae (B) / Fecal contamination of food and water / Humans
Cholera / Vibrio cholerae (B) / Fecal contamination of food and water / Humans
Table-1. The table clearly explains about the causative agent, Sources and reservoirs of the disease
3.1. Course of the Disease
Cholera (also called Asiatic cholera) is an infectious disease of the gastrointestinal tract caused by the Vibrio cholerae bacterium. These bacteria are typically ingested by drinking water contaminated by improper sanitation or by eating improperly cooked fish, especially shellfish. Symptoms include diarrhea, abdominal cramps, nausea, vomiting, and dehydration. Death isgenerally due to the dehydration causing by the illness.
The disease proceeds in possibly three stages.
a)Invasion b)Collapse c)Reaction
4. THE LAW OF TOTAL PROBABILITY
Although there are many solutions to prevent diseases, finding the right area to apply the prevention measure with right inputs becomes the criterion. The Bayes’ theorem evaluates the reverse of conditionality of events; where the symptoms and the causative-agents are analyzed
and found with a reciprocal equivalence. The Table-1 describes the most probable symptoms that cause the epidemics. The fact that the person had a positive reaction to the test may be considered as our data to build the collocation pattern.
The conditional probability of the collocation is the probability that a neighbor-set explaining the features of existence of causative agent, infection sources, is a part of the global neighbor-set in the spatial domain for this epidemic application Given a spatial domain in a database view S, to measure the implication strength of a spatial feature in a collocation pattern, a participation ratio pr(C,f) has to be defined. A feature f has a participation ratio pr(C,f) in pattern C means whenever the feature f is observed, with probability pr(C,f), all other features in C are also observed in a neighbor-set. In spatial application domain, as there are no natural transactions, for a continuous space, a participation index is proposed to measure the implication strength of a pattern from spatial features in the pattern. For a collocation pattern C, the participation index PI(C) = min fЄC {pr(C, f)}. In other words, wherever a feature in C is observed, with a probability of at least PI(C), all other features in C can be observed in a neighbor-set. A high participation index value indicates that the spatial features in a collocation pattern likely show up together.
Let us consider a collocation pattern:
C: {cause of epidemic} {causative agent, infection sources}; in the nearby region with high probability.
5. PROBLEM
Detection of the Epidemic
The collocation rules are very useful in detecting the affected areas by finding the symptoms of a disease and influence of symptoms in a disease by using sample identifiers, the collocation can be explained as follows: Assuming firstly, the ‘b’ as the consequence of feature ‘a’ is developed, forms a first level of collocation, which is identified by a → b, secondly, if the consequence ‘c’ from the feature ‘b’ is developed, forms a collocation, which is identified by b → c. As ‘b’ already have an antecedent ‘a’, the consolidated version of collocation, {a, b} → c can be formed. If ‘c’ becomes another feature that can lead to the consequence of ‘d’, then the notation wholly represents the cause of ‘d’ as {a, b, c} → d. Also implies to {a b c} → d representation.
Similarly, considering the collocation pattern for the problem:
C: {cause of epidemic} {causative agent, infection sources}; in the nearby region with high probability.
The collocation pattern is considered with practically proved parameters for cholera as follows …
As the disease reaches “Collapse” stage, the circulation is almost completely arrested, accelerated respiration, weak pulse, decreased systolic blood pressure, diminished or no urine output.
Assuming X as defined representation of collocated sequence of patterns i.e., C1, C2, C3, C4 the resultant collocation C5 is determined.. If the lead feature of the collocation contains higher probability then collocation is considered as highly important.The probabilities mentioned in the problem are <excreted along with innumerable Vibrios>, <loss of fluid, electrolyte imbalance>. If one of them or some of them exhibit high probability, then there is a high significance of occurring the disease severely, for low exhibition of probability, the existing of the disease will be indicative.
6. ALGORITHM
The following algorithm is to find the spatial knowledge i.e. cholera disaster from health demographic data.
1. Data collection from the patients.
2. Attributes related to cholera are selected.
3. Collocation rule is applied.
4. Spatial predicate is applied.
5. Source (Area) of disaster identified.
Synthesized data is generated using the rand () function. Parameters of cholera are chosen to work out from database.
Table-2: patient’s information
6.1 RESULTS
With threshold Value of 40%
Table-3: cholera effected areas
From the above result, we can say that near spatial object
(Area) “A10” people are having chances to get cholera.
7. NETWORK ARCHITECTURE
The built of framework explains the elements of the spatial knowledge support system in a work flow strategy and component architecture strategy .
Figure-2: Components of Frame work
The above conceptual design contains two important components, the spatial data mining infrastructure and the health campaign framework. The figure shows the detail process design framework of work flow strategy. The collocation pattern formed by this sample region acts as a cautious measure or the forecast for the bio-medical researchers, analysts and other health-care-takers of the spatial zone which will be useful for them to take suitable remedial campaigns.
Figure-3:Web-based Architecture for disaster Management
In Figure-3 web–based architecture was proposed for processing epidemic. DSL’s are nothing but stations across different locations, connected to central server through web.
Figure-4: Architecture of Epidemic System
Figure - 4 explains data flow of the system. In figure-5 proposed Client Server Architecture was given for the epidemic system .
Figure-5: Client-Server Software Architecture and its interacting components [Courtesy from Rainbow Architecture]
Socio-statistical methods related to health-science can be implemented to regulate the input variables that play a parametric role of collocation rule formation, in order to prevent the epidemic in the spatial zone, if not permanently, at least suitable preventive measures can be undertaken for the affect of such candidate epidemic in the interested spatial zone.
8. CONCLUSION
Epidemics, chronic diseases which are the major social disasters follow strategic-virulent disasters that affect the ecosystem of a spatial zone probabilistic study is made on the health demographic data. A Collocation rule is defined as a syntactic representation of the parameters in the form of antecedent and consequent. Using the Collocation rule, the effected area of cholera is found and results are obtained. Framework is described for the application of collocation rules i.e., spatial knowledge for the health campaign. Network architecture for spatial knowledge i.e. to identify cholera was proposed.
10. REFERENCES
[1] Chawla, Shekkar, Spatial Databases: A Tour, 2003, Prentice all. (ISBN – 013 – 017 480 – 7)
[2] Huang, Shekkar, Xiong, Discovery Collocation Patterns from Spatial Data Sets: A General Approach, IEEE-KDE,volume 16,No:12, dec 2004.
[3] Knorr and Ng, Extraction of Spatial Proximity Patterns by Concept Generalization, proceedings second international conference of KDD, pp. 347 – 350, aug 1996.